Download indirect interactions mediated by changing plant chemistry: beaver

ECOLOGY
Friday Sep 18 03:03 PM
Allen Press • DTPro
ecol 79
118 Mp
192 v File # 18sc
Ecology, 79(1), 1998, pp. 192–200
q 1998 by the Ecological Society of America
INDIRECT INTERACTIONS MEDIATED BY CHANGING PLANT
CHEMISTRY: BEAVER BROWSING BENEFITS BEETLES
GREGORY D. MARTINSEN, ELIZABETH M. DRIEBE,
AND
THOMAS G. WHITHAM
Department of Biological Sciences, Northern Arizona University, Box 5640,
Flagstaff, Arizona 86011 USA
Abstract. We documented an indirect interaction between beavers ( Castor canadensis)
and leaf beetles (Chrysomela confluens), mediated by changing plant chemistry of their
cottonwood hosts (Populus fremontii 3 P. angustifolia). Resprout growth arising from the
stumps and roots of beaver-cut trees contained twice the level of defensive chemicals as
normal juvenile growth. However, rather than being repelled by these defenses, leaf beetles
were attracted to resprout growth, resulting in a strong positive association between beavers
and beetles. Why? Cottonwoods contain phenolic glycosides, chemicals that are defensive
against mammalian herbivores, but are sequestered and used by the beetles for their own
defense. Experiments showed that beetles fed resprout growth were better defended against
their predators than those fed nonresprout growth. There also may have been a nutritional
benefit, because the conversion of the plant’s defense, salicin and other phenolic glycosides,
to salicylaldehyde releases glucose. Also, resprout growth contained more total nitrogen
than did nonresprout growth. Transfer experiments showed that, in apparent response to
these increased nutritional benefits, beetles fed resprout growth developed faster and
weighed more at maturity.
Although indirect interactions are much less studied than direct interactions, our work
suggests that the indirect interactions resulting from beaver cutting of cottonwoods have
important consequences for other organisms and could represent an important component
of community structure. The habitat mosaics created by beaver herbivory increase arthropod
biodiversity and may benefit other organisms such as birds and mammals. Furthermore, by
stimulating the production of resprout growth, beavers may play an important role in the
regeneration of a habitat type that is rapidly vanishing in the West.
Key words: beavers; cottonwoods; defensive chemistry; indirect interactions; leaf beetles; phenolic glycosides; sequestration.
INTRODUCTION
Through its feeding and dam-building activities, the
beaver, Castor canadensis, acts as a keystone species
and has disproportionate impacts on the community
(Naiman et al. 1986, Johnston and Naiman 1990, Jones
et al. 1994, Naiman et al. 1994). Once a major herbivore of riparian habitats, beavers were heavily trapped
and extirpated from most of North America by the turn
of the century (Jenkins and Busher 1979). In recent
years, however, they have returned to much of their
former range. The removal of dominant herbivores
such as beavers can have profound impacts on the landscape. For example, the near-elimination of beavers
from Yellowstone National Park in the 1800s is thought
to have resulted in the conversion of beaver ponds and
riparian habitat to dry grasslands (Chadde and Kay
1991). Here, we examine how beaver feeding impacts
an abundant insect herbivore and affects a habitat type
that is designated as threatened due to human activity
(Noss et al. 1995).
Few studies have examined indirect interactions beManuscript received 7 October 1996; revised 28 February
1997; accepted 1 March 1997.
tween distantly related taxa. More common are studies
of distant relatives that interact directly, and these interactions are often negative, e.g., competition (Brown
and Davidson 1977, Hay and Taylor 1985) and predation (Paine 1966, Estes and Palmisano 1974). However, indirect interactions in which one organism benefits from the actions of another may be common; such
interactions are not well documented because they are
probably more difficult to detect. For example, in the
short term, desert ants, birds, and mammals compete
for seeds (Brown et al. 1986), but in the long term,
rodents facilitate both ants (Davidson et al. 1984) and
birds (Thompson et al. 1991).
Many studies have shown that herbivores affect
plants and plant communities (reviewed by Huntley
1991), but few have examined the effects of one herbivore on other herbivore species. Mammalian herbivores are notable because they impact plant communities in two major ways. First, their browsing impacts
vegetation and can change competitive interactions
among dominant species (McNaughton 1976, Cantor
and Whitham 1989, McInnes et al. 1992). Second, the
disturbances they create greatly alter landscapes, e.g.,
beaver dams (Naiman et al. 1986), mima mounds
192
ECOLOGY
Friday Sep 18 03:03 PM
Allen Press • DTPro
January 1998
ecol 79
BEAVER BROWSING BENEFITS BEETLES
(Mielke 1977), and other soil disturbances created by
fossorial mammals (Platt 1975, Reichman and Smith
1985, Huntley and Inouye 1988, Whicker and Detling
1988, Martinsen et al. 1990). Beavers are especially
important because they radically alter communities in
both ways. The changes in plant communities that
beavers create almost certainly affect other herbivores
and other trophic levels.
At the individual plant level, it is well known that
herbivory can induce changes in defensive chemistry
(e.g., Tallamy and Raupp 1991). Such changes are likely when beavers cut down cottonwoods (Populus sp.),
removing nearly all aboveground biomass. When felled
by beavers, these trees often vigorously resprout from
their stumps and sucker from their roots. We predicted
that resprout growth resulting from beaver herbivory
would contain higher levels of defensive chemicals
(Basey et al. 1990). Phenolic glycosides are a group of
defensive chemicals found in many members of the
Salicaceae (Palo 1984). Although higher levels of phenolic glycosides in resprout growth could repel generalist herbivores, here we specifically examine how
these plant defenses might attract a specialist insect
herbivore, the leaf beetle Chrysomela confluens, which
is known to sequester them for its own defense (Kearsley and Whitham 1992). If beaver herbivory causes
an increase in plant defensive chemicals, then leaf beetles benefit.
Chrysomela confluens (Coleoptera Chrysomelidae)
is a chemically defended leaf beetle that feeds predominately on juvenile cottonwood trees (Kearsley and
Whitham 1989, Floate et al. 1993). We studied these
beetles and their cottonwood hosts (Populus fremontii,
P. angustifolia, and their hybrids) along the Weber River in northern Utah, USA. Adult beetles emerge in
April, feed on developing leaves, mate, and lay eggs
on the leaves. One female may produce several clutches
of 10–40 eggs. The larvae skeletonize cottonwood
leaves often completely defoliating small trees (Floate
and Whitham 1994). As they feed on leaf tissue, larvae
consume salicin and other phenolic glycosides that
break down to salicin. Ingested salicin is transported
to paired dorsal glands, where it is converted to salicylaldehyde (Pasteels et al. 1983). When a larva is
threatened, it everts these glands, exposing the potential predator to the salicylaldehyde. All larval development is completed by mid-July. The new adults feed
briefly before dropping to the ground, where they overwinter.
Beavers are locally abundant on the Weber River and
are important cottonwood herbivores. They cut down
trees ranging in size from seedlings to 1 m diameter
mature trees. When cut down or otherwise disturbed,
most of the cottonwoods at our study sites resprouted.
Our initial observations showed that beetles appeared
to be concentrated on resprout growth. We hypothesized that the resprout growth is better beetle habitat
than comparably sized young cottonwoods for two rea-
118 Mp 193
File # 18sc
193
sons: (1) resprout growth is likely to be better defended
and to contain more phenolic glycosides, which are the
basis of the beetles’ defense; and (2) resprout growth
is a better food resource, as it is more vigorous and
probably higher in limiting nutrients such as nitrogen.
METHODS
Beaver/beetle association
In order to document the relationship between
beavers and beetles, specifically, that beetles prefer cottonwood resprouts caused by beaver herbivory, we censused adult beetles on resprout (cottonwood ramets
arising from a stump) and adjacent nonresprout (other
juvenile cottonwood ramets) growth. These censuses
were conducted on hybrid cottonwoods, and the ramets
were matched for height. We performed a total of 22
paired counts, each 2 min long. The data were analyzed
using a paired t test.
Where beavers are active, they commonly cut down
several cottonwoods, creating patches of potential
beetle habitat. To determine whether beetles respond
positively to increased patch size, we set up circular
plots 4 m in diameter and counted number of beavercaused stumps, number of resprout ramets, and number
of clutches of Chrysomela confluens larvae. We censused a total of 25 of these plots. The data were analyzed using regression analysis.
Assays of beetle performance
If resprout cottonwood caused by beaver herbivory
contains more phenolic glycosides, then C. confluens
larvae feeding on resprout foliage should be better defended than those feeding on nonresprout growth. To
test this prediction, we collected larvae from .100
different ramets of both types of foliage and exposed
them to thatching ant (Formica propinqua) predators.
This bioassay is reasonable, because beetle larvae are
commonly encountered by ants tending colonies of
free-feeding aphids (Chaetophorus sp.) on juvenile cottonwood trees. In the field, there are as many as five
ant–beetle encounters per minute (Kearsley and Whitham 1992, Floate and Whitham 1994). We placed pairs
of larvae on ant mounds and recorded the time it took
for the defenses of each larva to be overwhelmed, at
which time they were dragged into the mound. The
larvae were always paired for size. We replicated this
experiment using six different ant mounds. The data
were analyzed using paired t tests.
Cottonwoods are easily cloned, and we have established common gardens by vegetatively propagating
known genotypes of cottonwoods and then planting
several copies of each genotype at the same site. At
the time of the study, these gardens were 10 years old
and included clones of one Fremont cottonwood, three
narrowleaf cottonwoods, and 15 hybrids. Because
beavers have cut down several trees, our gardens contained pairs of resprout and nonresprout trees from the
ECOLOGY
Friday Sep 18 03:03 PM
Allen Press • DTPro
ecol 79
GREGORY D. MARTINSEN ET AL.
194
same clone. We used these gardens to (1) conduct bioassays using C. confluens and (2) compare leaf chemistry of resprout and nonresprout cottonwoods, while
holding both tree genotype and site constant.
To determine whether resprout cottonwood is a better
food resource for leaf beetles, we experimentally transferred newly hatched first-instar larvae of C. confluens
to the same clones of resprout and nonresprout cottonwoods. To control for genetic differences among beetles, we split two clutches of larvae: half of each clutch
was transferred to resprout foliage and half to nonresprout foliage. Thus, we were able to compare the performance of beetles that were at least half siblings. We
transferred 30–35 larvae onto each tree type and then
placed cages over the shoots. The cages always included enough foliage to allow the larvae to complete
their development. Chrysomela confluens passes
through three larval instars and then pupates on or near
the host plant (in this experiment, many of the pupae
attached themselves to the inside of the cage). We collected the pupae, placed them in petri dishes, and waited for the new adults to emerge. We separated the beetles by sex and recorded: (1) time in days to adult
emergence and (2) mass of the newly emerged adults.
For each tree, we computed means for male and female
development time and mass. These data (four measures
of beetle performance for 18 pairs of resprout and nonresprout trees) were analyzed using a one-way MANOVA, with tree genotype as a blocking factor, followed
by paired t tests with sequential Bonferroni adjustments
(Rice 1989).
Leaf biochemistry
We collected leaves from pairs of resprout and nonresprout cottonwoods growing in our common gardens
in June 1992. This was the time when densities of larval
C. confluens on naturally occurring cottonwoods were
highest. We collected only the phenologically youngest
leaves. Because resprout leaves are much larger, we
had to sample from more nonresprout shoots to obtain
an equivalent amount of young leaf material. The
leaves were immediately frozen using dry ice, transported to an ultracold (2708C) freezer, and later freezedried in preparation for chemical analyses. Our experiments indicate that this is the best method for preventing interconversion of phenolic glycosides (E. M.
Driebe, unpublished data).
We measured concentrations of phenolic glycosides
in cottonwood leaves using high-performance liquid
chromatography (HPLC). Our protocol was a modification of the one described by Lindroth and Pajutee
(1987). Because other phenolic glycosides (e.g., salicortin, tremuloidin, tremulacin) can be converted to
salicin (Lindroth and Pajutee 1987, Clausen et al.
1989), we also measured their concentrations in the
resprout and nonresprout samples. Individual phenolic
glycoside concentrations were quantified using stan-
118 Mp 194
File # 18sc
Ecology, Vol. 79, No. 1
FIG. 1. Timed, 2-min censuses of adult Chrysomela confluens (mean 1 1 SE) on resprout and adjacent nonresprout
cottonwood foliage show that beetles are 15 times more abundant on resprout growth (n 5 22, paired t 5 7.28, P , 0.001).
Cottonwoods felled by beavers frequently resprout from the
stumps and roots.
dards. These defensive chemistry data were analyzed
with paired t tests.
The conversion of salicin to salicylaldehyde by
chrysomelid larvae liberates a molecule of glucose,
which is then available to the larva for nutrition (Pasteels et al. 1983, 1990). This sequestration may represent no metabolic cost, because the beetles derive
both defense against predators and nutrition from the
host plant defenses (Kearsley and Whitham 1992). It
may even represent a net benefit to the beetle if the
increased performance is a result of their ingesting
more plant defenses. However, because many studies
demonstrate a positive relationship between plant nitrogen concentration and insect performance (reviewed
by Mattson 1980, Scriber and Slansky 1981), we also
measured total nitrogen in our resprout and nonresprout
leaves.
Total nitrogen and total carbon were measured using
a LECO CNS-2000 (LECO, St. Josephs, Michigan,
USA). Approximately 1 g of freeze-dried tissue was
completely combusted at 13508C. Carbon was measured using an infrared detector; the nitrogen detector
measured thermal conductivity. Data were analyzed using paired t tests.
RESULTS
Beavers and beetle distribution
At two levels of examination, we found a close association between beavers and leaf beetles. First, the
distribution of adult beetles is strongly associated with
resprout cottonwood foliage arising from beaver herbivory. There were 15 times more adult beetles on resprout growth than on adjacent nonresprout growth
(Fig. 1). Beetles are probably attracted to these vigorously growing plants because of both reduced leaf
toughness and increased nutritional quality. An alter-
ECOLOGY
Friday Sep 18 03:03 PM
Allen Press • DTPro
January 1998
ecol 79
BEAVER BROWSING BENEFITS BEETLES
FIG. 2. The abundance of the leaf beetle C. confluens is
strongly associated with the amount of resprout growth
caused by beaver felling of cottonwoods (y 5 1.14 1 11.05x,
r2 5 0.84, P , 0.005). Resprout ramets were highly correlated
with stumps and were dropped from the multiple regression
model. Censuses were conducted on 4 m diameter circular
plots. Increased numbers of stumps generated more resprout
cottonwood ramets and more beetle habitat.
native explanation is that beetle predators are more
abundant on nonresprout growth and that resprout
growth represents enemy-free space. However, this
possibility seems unlikely; the arthropod and vertebrate
predators examined at our study sites have been shown
to forage in a density-dependent fashion (Dickson and
Whitham 1996).
Second, increased beaver activity generates more resprout ramets for more beetles to feed upon. Not only
are adult beetles more abundant on resprout growth,
but they also respond positively to patch size, laying
more clutches of eggs where stump density and resprout growth are greatest (Fig. 2). Eighty-four percent
of the variation in egg/larval clutch number is explained by the number of stumps. There are also strong
relationships between number of stumps and number
of resprouts (r2 5 0.72), and number of resprouts and
clutches of beetle larvae (r2 5 0.63). Because we censused clutches of young larvae or eggs prior to a buildup of predators, these results were little affected by
predation. Thus, increased beetle densities are most
likely due to their preference for large patches of resprout cottonwood resulting from beaver herbivory.
118 Mp 195
File # 18sc
195
muloidin, and tremulacin are labile compounds, and
their hydrolytic degradation produces salicin (Lindroth
and Pajutee 1987). In fact, crushing of leaf tissue, designed to mimic the effects of chewing insects, resulted
in the complete conversion of salicortin to salicin
(Clausen et al. 1989). Chrysomelid beetles then use the
salicin for both defense (salicylaldehyde) and nutrition
(glucose). Thus, higher levels of phenolic glycosides
in resprout cottonwood benefit these beetles in two important ways.
Although there were no differences in total carbon
between leaves of resprout and nonresprout ramets
from the same clone, levels of total nitrogen were almost 20% higher in resprout leaves (Fig. 4). This increased plant nitrogen probably provided an additional
nutritional benefit to the developing beetle larvae (see
the following section).
Beetle performance
Not only do beetles show a strong association with
resprout cottonwood (Figs. 1 and 2), but also their performance on resprout foliage is better than on nonresprout growth. Development times on resprout growth
are significantly faster (10%, on average), and adult
mass is significantly (16–20%) greater (Fig. 5). Growth
rate, the combination of development time and adult
mass, on resprout foliage was 30–35% greater than on
nonresprout foliage. These data support the hypothesis
that resprout cottonwood is a better food resource. As
mentioned previously, both high levels of phenolic glygosides and high levels of nitrogen in cottonwood
leaves could provide a nutritional benefit to the beetles.
Phytochemistry
Consistent with the hypothesis that beetles are attracted to greater amounts of defensive chemicals in
resprout foliage, we found that leaves of resprout cottonwood ramets had twofold higher phenolic glycoside
levels than those of nonresprout ramets (Fig. 3). For
all comparisons, both site and tree genotype were held
constant; in most cases, these were the same trees that
were used in the performance experiments previously
described. Salicortin, not salicin, was by far the most
abundant phenolic glycoside. However, salicortin, tre-
FIG. 3. Resprout cottonwood leaves contain, on average,
twice the concentration (mean 1 1 SE) of phenolic glycosides
as nonresprout leaves (n 5 15, paired t 5 3.31, P , 0.05).
The total includes salicin, salicortin, tremulacin, and tremuloidin. Salicortin is by far the most common phenolic glycoside. Larval beetles sequester these plant defensive chemicals and use them for both nutrition and their own defense.
ECOLOGY
Friday Sep 18 03:03 PM
Allen Press • DTPro
ecol 79
GREGORY D. MARTINSEN ET AL.
196
118 Mp 196
File # 18sc
Ecology, Vol. 79, No. 1
venile and mature shoots. The concentrations of two
chemicals that are breakdown products of phenolic glycosides were highest in juvenile Populus balsamifera
(Reichardt et al. 1990). Additionally, Tahvanainen et
al. (1985a) found much higher concentrations of phenolic glycosides in juvenile twigs than in mature twigs
of several willows (Salix sp.)
Recent evidence suggests that the breakdown products of phenolic glycosides, not the glycosides themselves, are defensive toward insect and mammalian herbivores (Clausen et al. 1989, Reichardt et al. 1990).
Salicortin, the most abundant phenolic glycoside in our
cottonwoods, is converted to salicin and 6-hydroxy-2cyclohexenone (6-HCH). Clausen et al. (1989) showed
that 6-HCH is an electrophile, capable of reacting with
biological nucleophiles such as proteins and nucleic
acids. Under basic conditions (e.g., insect gut), 6-HCH
converts to catechol, a substance toxic to insects (Todd
et al. 1971, Reese and Beck 1976, Clausen et al. 1989).
Under acidic conditions, as are found in the mammalian
gut, 6-HCH converts to phenol, which is highly toxic
FIG. 4. Resprout and nonresprout leaves contain equal
percentages (mean 1 1 SE) of total carbon (n 5 15, paired t
5 1.63, P . 0.05). However, resprout leaves are almost 20%
higher in total nitrogen (n 5 15, paired t 5 4.12, P , 0.001),
which is potentially very important to beetle nutrition.
Consistent with the hypothesis that beetles are attracted to highly defended resprout growth, where they
increase their own levels of defensive chemicals, we
found that larvae that feed on resprout cottonwood foliage survive an average of 28% longer when attacked
by ants than do larvae fed nonresprout leaves. This
relationship is statistically significant (P , 0.05) in five
of six sets of experimental trials (Fig. 6). These results
strongly suggest that beetle larvae feeding on resprout
cottonwood acquire more phenolic glycosides from
their host plant. Ants are a major predator of chrysomelid beetle larvae (Floate and Whitham 1994), and
thus represent a realistic bioassay of the effectiveness
of larval defenses against predation.
DISCUSSION
Defensive chemistry of phenolic glycosides
Phenolic glycosides are effective against mammalian
herbivores (Markham 1971, Edwards 1978, Tahvanainen et al. 1985a, Reichardt et al. 1990). Concentrations of these compounds can vary within a plant; in
particular, they are highest in juvenile tissues that are
most subject to browsing. Hares (Lepus spp.) prefer to
feed on mature salicaceous trees, avoiding juvenile
trees; within trees, they avoid juvenile shoots such as
those generated by previous herbivory (Bryant 1981,
Bryant et al. 1983, Tahvanainen et al. 1985a, Reichardt
et al. 1990). Basey et al. (1990) found similar preferences for beavers feeding on quaking aspen (Populus
tremuloides). Many of these studies have also shown
differences in defensive phytochemistry between ju-
FIG. 5. Chrysomela confluens larvae develop 10% faster
and grow 20% larger when feeding on resprout cottonwood
foliage, based on a sample of 30–40 larvae that were transferred onto 18 pairs of resprout and nonresprout cottonwoods.
Data shown are means 1 1 SE, and all differences are statistically significant (P , 0.05). The data are analyzed separately
by sex, because female beetles are 25–30% larger than males.
ECOLOGY
Friday Sep 18 03:03 PM
Allen Press • DTPro
January 1998
ecol 79
BEAVER BROWSING BENEFITS BEETLES
118 Mp 197
File # 18sc
197
FIG. 6. Results of paired trials comparing
survival of C. confluens larvae from resprout
and nonresprout foliage when placed on ant
mounds show that larvae are better defended
when fed resprout growth. Each trial consisted
of ;20 pairs of larvae placed on an ant mound,
one pair at a time. Survival times are the time
(mean 1 1 SE) it took for larvae to be dragged
into the ant mound. Asterisks denote significant
differences (P , 0.05) using paired t tests.
to mammals (Sax and Lewis 1989). For example, as
little as 1 g of phenol can kill a human (Budavari 1989).
Sequestration of an induced defense
Chrysomelids that are specialized to feed on willows
and cottonwoods (Salicaceae) incorporate salicin from
the host plant and convert it to salicylaldehyde, which
is an effective defense against arthropod predators such
as ants (Pasteels et al. 1983, Kearsley and Whitham
1992). Salicin and related phenolic glycosides are not
effective defenses against these specialist insects; in
fact, the beetles benefit from ingesting salicin, which
they convert to salicylaldehyde and glucose. However,
it is not known how beetle larvae are able to deal with
the other breakdown products of salicortin, 6-HCH and
catechol, which are toxic.
Although beetles sequester plant defenses, the increased production of phenolic glycosides by resprout
cottonwood ramets is probably an induced defense
against mammalian herbivores such as beavers (Basey
et al. 1990). Concentrations of these compounds can
vary within a plant; in particular, they are highest in
juvenile tissues such as those generated by previous
herbivory (Tahvanainen et al. 1985a, Basey et al. 1990,
Reichardt et al. 1990). Some authors argue that because
mammalian herbivory causes a phase change from mature to juvenile growth, it is not an example of an
induced response (Karban and Myers 1989, Bryant et
al. 1991). However, because we found more defensive
compounds in resprout growth than in other juvenile
growth, induction is suggested. Another possibility is
that the difference between resprout growth and unbrowsed juvenile trees is simply a consequence of the
relaxation of juvenile-phase defenses in the unbrowsed
trees, a process that occurs in as little as 2–3 years in
Populus (Fox and Bryant 1984).
Chrysomela confluens both prefers resprout cottonwood foliage and performs better when fed resprout
foliage. Several other studies have shown that chrysomelids feeding on salicaceous hosts prefer trees with
high leaf phenolic glycoside concentrations (Rowell-
Rahier 1984, Smiley et al. 1985, Tahvanainen et al.
1985b, Denno et al. 1990, Rank 1992, 1994). Smiley
et al. (1985) documented lower levels of natural predation on beetles feeding on willows with high phenolic
glycoside levels. Denno et al. (1990) showed mixed
results: beetles fed willows rich in phenolic glycosides
were better defended against coccinellid predators, but
larval development time and pupal dry mass were not
different for larvae reared on willows with high levels
vs. low levels of glycosides. Rank (1994) found only
a weak relationship between host plant chemistry and
larval survival, probably because the specialist predators in his system were not affected by larval defensive
secretions. In our field assays, beetle larvae feeding on
leaves highest in phenolic glycosides (resprout growth)
developed faster, were larger at maturity, and were better defended against generalist predators.
Indirect interactions
An indirect interaction occurs when the impact of
one species on another requires the presence of a third
species (Wootton 1994). Experimental ecological research has been dominated by studies of direct interactions (e.g., competition, predation, herbivory), and
the potential importance of indirect interactions has
only recently been appreciated (Carpenter 1988,
Strauss 1991, Wootton 1994).
Many of the best illustrations of indirect interactions
involve predators. For example, the classic studies of
keystone species (Paine 1966, 1969) are of keystone
predators. Indirect interactions are also better documented in aquatic systems than in terrestrial ones. One
common type of indirect interaction, trophic cascades,
in which the top-down effects of a predator suppress
herbivores and benefit plants, are largely unknown in
terrestrial systems (Strong 1992; but see Marquis and
Whelan 1994 for an exception). In terrestrial systems,
perhaps it is better to look for indirect interactions
among herbivores, their host plants, and other herbivore species. Several examples include closely related
ECOLOGY
Friday Sep 18 03:03 PM
Allen Press • DTPro
198
ecol 79
GREGORY D. MARTINSEN ET AL.
herbivore taxa (McNaughton 1976, Faeth 1986, Hunter
1987, 1992).
We have shown that beaver herbivory facilitates leaf
beetle growth rates and defense by changing the chemistry of their cottonwood hosts. Positive indirect interactions between distantly related herbivores may be
common, even in terrestrial systems, but they are difficult to detect. Granivorous desert rodents facilitate
both ants (Davidson et al. 1984) and birds (Thompson
et al. 1991), and winter browsing of birch by moose
later benefits several species of insects (Dannell and
Huss-Dannell 1985). Mammalian herbivores that can
alter plant growth form, phytochemistry, and even species composition are likely to interact indirectly with
a wide variety of organisms and ultimately impact community structure, in much the same way that top predators control some aquatic communities.
Beavers and riparian areas
Although cottonwood trees are one of the most productive and sensitive components of riparian ecosystems in western North America, this habitat type is
vanishing rapidly: .100 000 ha are lost annually (Finch
and Ruggiero 1993). Cottonwoods support many species, both invertebrates and vertebrates, especially
birds (Carothers et al. 1974, Strong and Bock 1990).
Beavers are thought to be important agents in the formation and preservation of riparian areas (Naiman et
al. 1986, 1994, Chadde and Kay 1991), and beaver
herbivory may be very important to the maintenance
of cottonwood stands.
With the advent of dam building and the resultant
decline in natural flooding, which is crucial for creating
seed beds for cottonwood establishment (Rood and Mahoney 1990), cottonwoods that typically resprout after
beaver cutting may ultimately be dependent on beavers
for regeneration. Furthermore, species such as Chrysomela confluens that are restricted to juvenile cottonwoods may become increasingly dependent on the resprout growth caused by beavers.
Beaver herbivory creates a habitat mosaic of mature
trees and juvenile resprout growth. These mixed-aged
stands promote biodiversity, because juvenile and mature ramets of the same cottonwood clone support different arthropod communities (Waltz and Whitham
1997). Beaver herbivory probably also benefits vertebrates. For example, mixed-aged stands support greater
avian biodiversity than even-aged stands (Derleth et al.
1989).
Beavers are considered keystone species that can literally alter the environment at the landscape level.
Here, we have shown their diverse impacts on plant
chemistry and on an insect that sequesters plant defenses for its own use. Because the sequestered chemicals may also be used for nutrition, this is an example
of a no-cost defense. Beaver herbivory could be important in affecting the geographic distribution of the
beetles and stimulating regeneration of cottonwoods,
118 Mp 198
File # 18sc
Ecology, Vol. 79, No. 1
which are currently suffering widespread decline in the
West.
ACKNOWLEDGMENTS
We thank J. Bryant, K. Charters, N. Cobb, J. Coleman, K.
Floate, J Ganey, K. Gehring, J. Pasteels, P. Price, T. Theimer,
K. Whitham, and an anonymous reviewer for comments on
the manuscript. S. Denton, L. Dickson, A. Martin, and B.
Strickland assisted with the field work. We thank T. Clausen,
D. Gilbert, R. Lindroth, and D. A. Wait for suggestions on
HPLC, T. Clausen for providing phenolic glycoside standards,
and R. Turek for statistical advice. This work was supported
by NSF grants DEB-9311210, DEB-9408009, and USDA
92-37302-7854.
LITERATURE CITED
Basey, J. M., S. H. Jenkins, and G. C. Miller. 1990. Food
selection by beavers in relation to inducible defenses of
Populus tremuloides. Oikos 59:57–62.
Brown, J. H., and D. W. Davidson. 1977. Competition between seed-eating rodents and ants in desert ecosystems.
Science 196:880–882.
Brown, J. H., D. W. Davidson, J. C. Munger, and R. S. Inouye.
1986. Experimental community ecology: the desert granivore system. Pages 41–61 in J. L. Diamond and T. J. Case,
editors. Community ecology. Harper and Row, New York,
New York, USA.
Bryant, J. P. 1981. Phytochemical deterrence of snowshoe
hare browsing by adventitious shoots of four Alaskan trees.
Science 313:889–890.
Bryant, J. P., F. S. Chapin III, and D. R. Klein. 1983. Carbon/
nutrient balance of boreal plants in relation to vertebrate
herbivory. Oikos 40:357–368.
Bryant, J. P., K. Danell, F. Provenza, P. B. Reichardt, T. A.
Clausen, and R. A. Werner. 1991. Effects of mammal
browsing on the chemistry of deciduous woody plants.
Pages 135–155 in D. W. Tallamy and M. J. Raupp, editors.
Phytochemical induction by herbivores. John Wiley, New
York, New York, USA.
Budavari, S., editor. 1989. The Merck Index. Merck, Rahway, New Jersey, USA.
Cantor, L. F., and T. G. Whitham. 1989. Importance of belowground herbivory: pocket gophers may limit aspen to
rock outcrop refugia. Ecology 70:962–970.
Carothers, S. W., R. R. Johnson, and S. W. Aitchison. 1974.
Population and social organization of southwestern riparian
birds. American Zoologist 14:97–108.
Carpenter, S. C., editor. 1988. Complex interactions in lake
communities. Springer-Verlag, New York, New York, USA.
Chadde, S. W., and C. E. Kay. 1991. Tall willow communities
on Yellowstone’s northern range: a test of the ‘‘natural
regulation’’ paradigm. Pages 231–262 in R. B. Keiter and
M. S. Boyce, editors. The Greater Yellowstone ecosystem.
Yale University Press, New Haven, Connecticut, USA.
Clausen, T. P., P. B. Reichardt, J. P. Bryant, R. A. Werner, K.
Post, and K. Frisby. 1989. Chemical model for short-term
induction in quaking aspen (Populus tremuloides) foliage
against herbivores. Journal of Chemical Ecology 15:2335–
2346.
Dannell, K., and K. Huss-Dannell. 1985. Feeding by insects
and hares on birches earlier affected by moose browsing.
Oikos 44:75–81.
Davidson, D. W., R. S. Inouye, and J. H. Brown. 1984. Granivory in a desert ecosystem: experimental evidence for indirect facilitation of ants by rodents. Ecology 65:1780–
1786.
Denno, R. F., S. Larsson, and K. L. Olmstead. 1990. Role
of enemy-free space and plant quality in host plant selection
by willow leaf beetles. Ecology 71:124–137.
Derleth, E. L., D. J. McAuley, and T. J. Dwyer. 1989. Avian
ECOLOGY
Friday Sep 18 03:03 PM
Allen Press • DTPro
January 1998
ecol 79
BEAVER BROWSING BENEFITS BEETLES
community response to small-scale habitat disturbance in
Maine. Canadian Journal of Zoology 67:385–390.
Dickson, L. L., and T. G. Whitham. 1996. Genetically based
plant resistance traits affect arthropods, fungi, and birds.
Oecologia 106:400–406.
Edwards, W. R. N. 1978. Effect of salicin content on palatibility of Populus foliage to opossum (Trichosurus vulpecula). New Zealand Journal of Science 21:103–106.
Estes, J. A., and J. F. Palmisano. 1974. Sea otters: their role
in structuring nearshore communities. Science 185:1058–
1060.
Faeth, S. H. 1986. Indirect interactions between temporally
separated herbivores mediated by the host plant. Ecology
67:479–494.
Finch, D. M., and L. F. Ruggiero. 1993. Wildlife habitats
and biological diversity in the Rocky Mountains and northern Great Plains. Natural Areas Journal 13:191–203.
Floate, K. D., M. J. C. Kearsley, and T. G. Whitham. 1993.
Elevated herbivory in plant hybrid zones: Chrysomela confluens, Populus, and phenological sinks. Ecology 74:2056–
2065.
Floate, K. D., and T. G. Whitham. 1994. Aphid–ant interaction reduces chrysomelid herbivory in a cottonwood hybrid zone. Oecologia 97:215–221.
Fox, J. F., and J. P. Bryant. 1984. Instability of the snowshoe
hare and woody plant interaction. Oecologia 63:128–135.
Hay, M. E., and P. R. Taylor. 1985. Competition between
herbivorous fishes and urchins on Caribbean reefs. Oecologia 65:591–598.
Hunter, M. D. 1987. Opposing effects of spring defoliation
on late-season oak caterpillars. Ecological Entomology 12:
373–382.
. 1992. Interactions within herbivore communities
mediated by the host plant: the keystone herbivore concept.
Pages 287–325 in M. D. Hunter, P. W. Price, and T. Ohgushi,
editors. Effects of resource distribution on animal–plant
interactions. Academic Press, San Diego, California, USA.
Huntley, N. 1991. Herbivores and the dynamics of communities and ecosystems. Annual Review of Ecology and
Systematics 22:477–503.
Huntley, N., and R. Inouye. 1988. Pocket gophers in ecosystems: patterns and mechanisms. BioScience 38:786–
793.
Jenkins, S. H., and P. E. Busher. 1979. Castor canadensis.
Mammalian Species 120:1–8.
Johnston, C. A., and R. J. Naiman. 1990. Browse selection
by beaver: effects on forest composition. Canadian Journal
of Forest Research 20:1036–1043.
Jones, C. G., J. H. Lawton, and M. Shachak. 1994. Organisms as ecosystem engineers. Oikos 69:373–386.
Karban, R., and J. H. Myers. 1989. Induced plant responses
to herbivory. Annual Review of Ecology and Systematics
20:331–348.
Kearsley, M. J. C., and T. G. Whitham. 1989. Developmental
changes in resistance to herbivory: implications for individuals and populations. Ecology 70:422–434.
Kearsley, M. J. C., and T. G. Whitham. 1992. Guns and
butter: a no-cost defense against predation. Oecologia 92:
556–562.
Lindroth, R. L., and M. S. Pajutee. 1987. Chemical analysis
of phenolic glycosides: art, facts, and artifacts. Oecologia
74:144–148.
Markham, K. R. 1971. A chemotaxonomic approach to the
selection of opossum-resistant willows and poplars for use
in soil conservation. New Zealand Journal of Science 14:
179–186.
Marquis, R. J., and C. J. Whelan. 1994. Insectivorous birds
increase growth of white oak through consumption of leafchewing insects. Ecology 75:2007–2014.
Martinsen, G. D., J. H. Cushman, and T. G. Whitham. 1990.
118 Mp 199
File # 18sc
199
Impact of pocket gopher disturbance on a shortgrass prairie
community. Oecologia 83:132–138.
Mattson, W. J. 1980. Herbivory in relation to plant nitrogen
content. Annual Review of Ecology and Systematics 11:
119–161.
McInnes, P. F., R. J. Naiman, J. Pastor, and Y. Cohen. 1992.
Effects of moose browsing on vegetation and litter of the
boreal forest, Isle Royale, Michigan, USA. Ecology 73:
2059–2075.
McNaughton, S. J. 1976. Serengeti migratory wildebeest:
facilitation of energy flow by grazing. Science 191:92–94.
Mielke, H. W. 1977. Mound building by pocket gophers
(Geomyidae): their impact on soils and vegetation in North
America. Journal of Biogeography 4:171–180.
Naiman, R. J., J. M. Melillo, and J. E. Hobbie. 1986. Ecosystem alteration of boreal forest streams by beaver (Castor
canadensis). Ecology 67:1254–1269.
Naiman, R. J., G. Pinay, C. A. Johnston, and J. Pastor. 1994.
Beaver influences on the long-term biogeochemical characteristics of boreal forest drainage networks. Ecology 75:
905–921.
Noss, R. F., E. T. LaRoe III, and J. M. Scott. 1995. Endangered ecosystems of the United States: a preliminary assessment of loss and degradation. Biological Report 28.
U.S. Department of the Interior, National Biological Service, Washington, D.C., USA.
Paine, R. T. 1966. Food web complexity and species diversity. American Naturalist 100:65–75.
. 1969. A note on trophic complexity and community
stability. American Naturalist 103:91–93.
Palo, R. T. 1984. Distribution of birch (Betula spp.), willow
(Salix spp.), and poplar (Populus spp.) secondary metabolites and their potential role as chemical defense against
herbivores. Journal of Chemical Ecology 10:499–520.
Pasteels, J. M., S. Duffey, and M. Rowell-Rahier. 1990. Toxins in chrysomelid beetles. Possible evolutionary sequence
from de novo synthesis to derivation from food plant chemicals. Journal of Chemical Ecology 16:211–222.
Pasteels, J. M., M. Rowell-Rahier, J. C. Braekman, and A.
Dupont. 1983. Salicin from host plant as precursor of salicylaldehyde in defensive secretion of chrysomeline larvae.
Physiological Entomology 8:307–314.
Platt, W. J. 1975. The colonization and formation of equilibrium plant species associations on badger disturbances
in a tallgrass prairie. Ecological Monographs 45:285–305.
Rank, N. E. 1992. Host plant preference based on salicylate
chemistry in a willow leaf beetle (Chrysomela aeneicollis).
Oecologia 90:95–101.
. 1994. Host plant effects on larval survival of a salicin-using leaf beetle Chrysomela aeneicollis Schaeffer
(Coleoptera: Chrysomelidae). Oecologia 97:342–353.
Reese, J. C., and S. D. Beck. 1976. Effects of allelochemics
on the black cutworm: effects of catechol, L-dopa, dopamine, and chlorogenic acid on larval growth, development,
and utilization of food. Annals of the Entomological Society of America 69:68–72.
Reichardt, P. B., J. P. Bryant, B. R. Mattes, T. P. Clausen, F.
S. Chapin, III, and M. Meyer. 1990. Winter chemical defense of Alaskan balsam poplar against snowshoe hares.
Journal of Chemical Ecology 16:1941–1959.
Reichman, O. J., and C. S. Smith. 1985. Impact of pocket
gopher burrows on underlying vegetation. Journal of Mammalogy 66:720–725.
Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223–225.
Rood, S. B., and J. M. Mahoney. 1990. Collapse of riparian
poplar forests downstream from dams in western prairies:
probable causes and prospects for mitigation. Environmental Management 14:451–464.
Rowell-Rahier, M. 1984. The food plant preferences of Phra-
ECOLOGY
Friday Sep 18 03:03 PM
Allen Press • DTPro
200
ecol 79
GREGORY D. MARTINSEN ET AL.
tora vitellinae (Coleoptera: Chrysomelidae). A. Field observations. Oecologia 64:369–374.
Sax, N. I., and R. J. Lewis, Sr. 1989. Dangerous properties
of industrial materials. Seventh edition. Van Nostrand,
Reinhold, New York, New York, USA.
Scriber, J. M., and F. Slansky, Jr. 1981. The nutritional ecology of immature insects. Annual Review of Entomology
26:183–211.
Smiley, J. T., J. M. Horn, and N. E. Rank. 1985. Ecological
effects of salicin at three trophic levels: new problems from
old adaptations. Science 229:649–651.
Strauss, S. Y. 1991. Indirect effects in community ecology:
their definition, study, and importance. Trends in Ecology
and Evolution 6:206–210.
Strong, D. R. 1992. Are trophic cascades all wet? Differentiation and donor control in speciose ecosystems. Ecology 73:747–754.
Strong, T. R., and C. E. Bock. 1990. Bird species distribution
patterns in riparian habitats in southeastern Arizona. Condor 92:866–885.
Tahvanainen, J., E. Helle, R. Julkunen-Tiitto, and A. Lavola.
1985a. Phenolic compounds of willow bark as deterrents
against feeding by mountain hare. Oecologia 65:319–323.
118 Mp 200
File # 18sc
Ecology, Vol. 79, No. 1
Tahvanainen, J., R. Julkenen-Tiitto, and J. Kettunen 1985b.
Phenolic glycosides govern the food selection pattern of
willow feeding leaf beetles. Oecologia 67:52–56.
Tallamy, D. W., and M. J. Raupp, editors. 1991. Phytochemical induction by herbivores. John Wiley, New York, New
York, USA.
Thompson, D. B., J. H. Brown, and W. D. Spencer. 1991.
Indirect facilitation of granivorous birds by desert rodents:
experimental evidence from foraging patterns. Ecology 72:
852–863.
Todd, G. W., A. Getahun, and D. C. Cress. 1971. Resistance
in barley to the greenbug, Schizaphis graminum. I. Toxicity
of phenolic and flavinoid compounds and related substances. Annals of the Entomological Society of America
64:718–722.
Waltz, A. M., and T. G. Whitham. 1997. Plant development
affects arthropod community structure: opposing impacts
of species removal. Ecology 78, in press.
Whicker, A. D., and J. K. Detling. 1988. Ecological consequences of prairie dog disturbances. BioScience 38:778–
785.
Wootton, J. T. 1994. The nature and consequences of indirect
effects in ecological communities. Annual Review of Ecology and Systematics 25:443–466.